By Geoffrey Chen
The approach to singularity is not just another stage in technical progress. It marks a deeper change in the structure of knowledge itself. For several centuries, modern society has worked on a basic assumption that knowledge must be centered on a subject. A human being perceives, understands, judges, and reflects, and this inner process gives knowledge both its source and its legitimacy. Science, education, professional life, and social organization have all depended on this assumption in one way or another. Knowledge was treated as something that only really existed when a subject genuinely knew.
The development of artificial intelligence, especially autonomous agents and systems that can participate in improving later systems, is beginning to unsettle that assumption. The issue is no longer whether machines can assist human thought. The more serious issue is that many effective cognitive results can now be produced without the full participation of a human subject who understands them from within. Models can infer, generate plans, make decisions, coordinate tasks, and increasingly contribute to the construction of the next generation of models. Under these conditions, the long-standing bond between knowledge and the human subject starts to loosen.
This is the central shift that After the Knowing Subject tries to identify. Knowledge does not necessarily require an inner, unified, self-transparent knower. For a long time, we treated knowing as an internal state, as though knowledge became real only when a subject consciously grasped an object. What matters more today is something else. What matters is the reliability of outputs, the stability of structures, and the verifiable performance of systems in the world. In other words, knowledge is moving away from the question of who understands what, and toward the question of what kind of system can consistently produce trustworthy results.
From this angle, the real meaning of the so-called singularity is not simply greater speed or stronger capability, and not merely the claim that AI can now create AI. Its deeper meaning is that human beings are facing, at scale, a form of cognition that can continue to expand without depending on the traditional human subject. Earlier technological revolutions were still, in broad terms, extensions of the hand. They improved tools, production, and physical reach. This time the change touches analysis, judgment, expression, and decision-making, which were once treated as the core powers of the subject itself. For the first time, human beings are forced to confront the possibility that they are not the indispensable center of knowledge.
That is why the reaction is so often anxiety rather than simple curiosity. On the surface, people worry about jobs, competition, and the devaluation of skills. But underneath that is a deeper instability in identity. Modern people have long tied their dignity to cognitive superiority. We assumed that what makes human beings distinct is not merely that they live, but that they understand, reason, create, and organize meaning. Once these capacities are no longer exclusive to human beings, and no longer even primarily carried by them in many settings, the question of what a human being is becomes sharp again in a way it has not been for a long time.
Many discussions of AI drift into exaggeration and panic because they still think inside the old framework. People continue to describe AI as a tool, and so they keep asking whether it is controllable, whether it will become uncontrollable, whether it will replace certain professions, and whether it will generate risks. Those questions matter, but they do not reach the deepest level. The more basic question is this. If cognitive authority is shifting from the interior of the subject to external systems, how should we redefine knowledge, responsibility, value, and the human place in the world?
Seen this way, the idea of a stop button begins to look less like a real solution and more like a leftover comfort from an older picture of the world. We still imagine the human being as the center standing outside the system and issuing commands, while technology is treated as a passive instrument. But the rise of autonomous systems shows that reality no longer fits that model. What we are dealing with now is less a clean division between subject and tool than a cognitive network composed of human beings, models, platforms, data, and feedback loops. In such a network, no single subject fully commands the whole process. What is weakening is not just our technical control, but the clarity of the question of who is in control at all.
This has direct consequences for society. Modern economic life has long rested on a simple assumption that human labor and human cognitive ability form the basis of value distribution. Those who possess skill and perform work are entitled to reward. But if more and more high-value cognitive activity can be carried out by automated systems, then the old logic of distribution loses its stability. The problem is not only that jobs disappear. It is that the age in which human cognitive capacity functioned as a scarce resource is beginning to end. Once that happens, education, careers, class mobility, and the narrative of self-making all start to lose the support that once made them coherent.
The most dangerous part is not that machines may become smarter than people. It is that society is still trying to manage a new reality with old institutions. The mode of knowledge production has changed, but legitimacy, distribution, and structures of dignity have not changed with it. The result is a widening break. A small number of actors who control models, compute, platforms, and data will control the new cognitive infrastructure. Meanwhile, most people still inhabit the old mythology of the subject while gradually losing their place in the new order. What we call anxiety is therefore not simply a matter of individual psychology. It is a structurally normal response to the erosion of subject-based status.
But the significance of After the Knowing Subject is not that it amplifies this anxiety. Its value lies in offering a more disciplined way to understand what is happening. It does not defend human exceptionalism in any simple form, and it does not mystify AI. Its claim is that knowledge has never really been the product of an isolated subject alone. Scientific communities, institutional systems, technical networks, and now model architectures have always carried a tendency toward depersonalization. AI has simply pushed that tendency to a point where it can no longer be ignored. It has made visible something that was always partly true, namely that knowing is not only a matter of conscious possession, but also a matter of function, structure, and the reliable production of results.
Once this is accepted, human beings no longer need to tie their meaning to the old position of being the sole true knower. What must be rebuilt is not human cognitive superiority, but the human understanding of existence itself. Human importance does not depend on preserving a monopoly over reasoning, memory, and calculation. It depends on the fact that human beings remain exposed to value, finitude, suffering, mortality, and the burden of meaning. A machine may generate answers, but that does not place it inside the human condition. It may complete tasks, but that does not mean it bears human vulnerability. For that reason, human value in the future cannot continue to rest on the claim that we calculate better than machines. It has to rest on how human beings live and take responsibility within a world where knowledge is no longer centered on the subject.
This also changes the role of philosophy. In the age of AI, philosophy is not important because it can slow down technology. It is important because it helps us let go of self-images that no longer hold. We should no longer treat the subject as the unquestioned origin of all knowledge and value, and we should no longer turn consciousness into a sacred final refuge. The more realistic task is to distinguish between what belongs to the old mythology of the subject and what really remains irreducible in human life. Understanding can be outsourced. Calculation can be surpassed. Even many forms of creativity can be absorbed into systems. But responsibility, suffering, choice, relation, and the search for meaning within a finite life do not simply disappear.
So when artificial intelligence moves toward what people call the singularity, the first implication is not that machines will rule humanity. A more accurate statement is that the modern myth of the cognitive subject is reaching its limit. Human beings will no longer stand as the unique central node in the universe of knowledge, but as one participant within a larger cognitive order. This will certainly bring loss, as well as major institutional and ethical disruption. But it does not have to be understood only as a disaster. It may also force us to abandon the narrow belief that rational or cognitive superiority is the whole basis of human worth.
If this age imposes a serious task on us, it is not to fight desperately to preserve the throne of the old subject. It is to ask how a new understanding of the human can still be built after the subject has been displaced from the center. The most important question is no longer whether machines have consciousness in some abstract sense. The more urgent question is whether human beings can reorganize dignity, institutions, and ways of living in a world where knowledge no longer depends on the traditional figure of the knower.
That is the deepest philosophical problem opened by AI. It is not a secondary issue attached to technology. It is a test of the entire modern way of understanding ourselves. If there is a singularity here, it is not first a singularity of computing power. It is a singularity in our idea of the subject. It forces us to admit that knowledge no longer naturally belongs to the subject, and that human beings no longer occupy the center of the world simply by possessing knowledge. If humanity still has a chance to mature under these conditions, that process has to begin there.